S5: Protein engineering using machine learning and synthetic genes for commercial applications. Maximizing efficiency and minimizing cost.

Nature’s exquisite enzymatic chemistry tends toward the highly selective, efficient, temperate, and environmentally benign. Industrial biocatalysis has the ability not only to identify organisms and their enzymes, but to drastically engineer biocatalysts for improved performance.

Engineering of biological processes, including biocatalysis, protein expression and regulation requires a tool to define distances and directions in mega-dimensional space. It also requires the ability to interrogate predefined locations in this space for functional information. DNA2.0 Inc has developed the ProteinGPS suite of technology tools to accomplish these goals. We use modern machine learning tools in conjunction with efficient denovo gene synthesis to design and develop algorithms that allow the quantification and optimization of any bio-based process, including biocatalytic processes and protein expression. The starting coordinates for the sequence space exploration is extracted from the ever growing public DNA sequence repositories. Because of the efficiency of the system, all optimization can be done directly for the real-world application, and not towards a HTP surrogate assay which may or may not overlap with the real-world assay.